Overview

An ERP AI chatbot is an intelligent virtual assistant that integrates with Enterprise Resource Planning (ERP) systems, enabling users to interact with business data through natural language conversations. This tutorial shows how to build two different ERP automation workflows with n8n.

What is an ERP AI Chatbot?

ERP AI chatbots are intelligent assistants that bridge the gap between users and complex ERP systems. They allow employees to query business data, create records, and automate workflows using natural language instead of navigating complex ERP interfaces.

• Enhanced user experience - natural language instead of complex UI
• Improved data accessibility - query business data conversationally
• Automated routine tasks - reduce manual data entry
• Quick insights generation - instant answers to business questions

Use Cases for ERP AI Chatbots

• Automatically log sales inquiries as opportunities
• Query sales pipeline and revenue data
• Generate reports and summaries
• Update customer records
• Check inventory levels
• Create purchase orders
• Track project status

Prerequisites

• n8n instance (cloud or self-hosted)
• Odoo account (or other ERP system)
• OpenAI API key
• Gmail account (for email automation)
• Basic understanding of ERP concepts

Workflow 1: Email-to-Opportunity Automation

This workflow automatically converts sales inquiry emails into Odoo opportunities with AI-generated summaries.

Step 1: Configure Gmail Trigger

Set up a Gmail Trigger node to monitor incoming emails. You can filter by label (e.g., 'Sales') to only process relevant messages.

{
  "pollTimes": {
    "item": [
      {
        "mode": "everyMinute"
      }
    ]
  },
  "labelIds": ["Label_Sales"],
  "format": "simple"
}

Step 2: AI Summarization

Use OpenAI to summarize the email and extract key information like project details, budget, and timeline.

System Prompt:
Write a concise summary of the following sales inquiry email. Extract:
- Project description
- Estimated budget (if mentioned)
- Timeline or deadline
- Key requirements
- Contact preferences

Email content: {{$json.body}}

Step 3: Create Odoo Opportunity

Automatically create a new sales opportunity in Odoo using the extracted information. Add the AI summary as internal notes.

{
  "resource": "opportunity",
  "operation": "create",
  "name": "{{$json.subject}}",
  "partner_id": "{{$json.fromEmail}}",
  "description": "{{$json.aiSummary}}",
  "stage_id": 1
}

Workflow 2: Interactive ERP AI Chatbot

Build an intelligent chatbot that can answer questions about your sales pipeline, opportunities, and business data.

Step 4: Extract ERP Context

Retrieve relevant data from Odoo (opportunities, customers, products) and format it as context for the AI agent.

// Retrieve opportunities from Odoo
const opportunities = $('Odoo').all();

// Format for AI context
const context = opportunities.map(opp => ({
  name: opp.json.name,
  value: opp.json.expected_revenue,
  stage: opp.json.stage_id,
  probability: opp.json.probability
}));

return { context: JSON.stringify(context) };

Step 5: Configure AI Agent

Set up an AI Agent with access to ERP data and tools for calculations and data queries.

System Message:
You are an ERP assistant with access to our sales pipeline data. You can:
- Answer questions about opportunities, revenue, and forecasts
- Calculate totals and averages
- Provide insights on sales performance
- Help users find specific opportunities or customers

Context: {{$json.context}}

Be concise, accurate, and helpful. Always base your answers on the provided context.

Step 6: Add Tools

Enhance the agent with tools like a calculator for computing revenue totals, averages, and forecasts.

• Calculator tool - for revenue calculations
• Database query tool - for dynamic data retrieval
• Chart generation tool - for visual insights
• Custom code tool - for complex logic

Step 7: Deploy Chatbot

Make the chatbot available to your team through hosted chat or embed it in your internal tools.

Cost Management Strategies

• Keep chatbot focused on specific use cases
• Set token limits for requests and responses
• Use smaller models (gpt-3.5-turbo) for simple queries
• Cache frequently accessed ERP data
• Implement rate limiting for users
• Monitor usage and optimize prompts

Best Practices

• Start with simple automation before building complex chatbots
• Clearly define what data the chatbot can access
• Implement proper access controls and authentication
• Test thoroughly with sample data before production
• Monitor chatbot accuracy and user satisfaction
• Regularly update context data for accuracy